While this doesnt help your accuracy problem, one good way to measure the
frequency of a micro controller is to make a program which basically
divides the clock by some value (i personally use 30) and outputs it to a
pin. This is better than measuring directly at the crystal, both because
the capacitance of the measuring instrument can, from what ive heard, load
the oscillator, and moreover, i think frequency counters are more accurate
at lower frequencies.

------------------------------------------------------------------------------
A member of the PI-100 Club:
3.1415926535897932384626433832795028841971693993751
058209749445923078164062862089986280348253421170679

On Wed, 30 Jun 1999, Andres Tarzia wrote:

> Hi all!
>
> I am having problems with an 20Mhz oscillator that does not want to
> oscillate at 20Mhz. Something like 20,080,000Hz.
>
> I am using the "usual" oscillator construction with two fixed caps to
> ground, one at each end of the crystal (120pf in this case).
>
> The problem is that the damn circuit is oscillating at slightly above 20Mhz,
> so I changed one of the fixed caps (the OSC1/ClkIN) for a variable cap. The
> only one I had was a 5-20pf part. I was able to reduce the oscillation
> frequency to 20,035,000Hz but no less than that.
>
> I am using a temperature-compensated frequencimeter for measurement (using
> the OSC2/ClkOUT pin) that is very exact.
>
> What can I do to reduce it further? I was thinking about changing the (now
> only) fixed cap for another one, but with LARGER or SMALLER capacitance?
> Could someone help me here?
>
> Thank you very much!
>
> Regards,
> Andres Tarzia
> Technology Consultant - SMART S.A.
> email: atarzia@smart.com.ar
>